Abstract
Purpose: Positron Emission Tomography with 18F-Fluoride offers the opportunity to image bone metastases and measure their response to therapy with high sensitivity. The PET/CT combination is more accurate in staging skeleton metastases compared to conventional imaging methods. Indeed 18F-Fluoride is a bone specific tracer that offers the opportunity to monitor osteoblastic activity, which can be also detected in the CT due to the resulting tissue mineralization. However, while CT is an imaging technique characterized by high resolution, the clinical potential of PET is affected by its low resolution and resulting partial volume effect. Here we optimize and evaluate a synergistic resolution recovery approach that combines functional (PET) and structural (CT) information. The resulting mitigation of partial volume effect will ultimately lead to more robust tumour staging.
Methods: The proposed method takes advantage of the multiresolution property of the wavelet transform that is therefore applied to both functional and structural images. The correspondent resolution components of PET and CT are then combined after appropriate scaling and weighting. Although the method was validated for PET/MRI brain data we demonstrated the potentiality of our modified version with both phantom and patient experiment. We additionally compared the method with a standard resolution recovery approach based on point-spread-function reconstruction.
Results: The activity quantification and volume definition on phantom spheres showed better performance of our methodology compared to the clinical standard and the point-spread-function reconstruction. The used of segmented CT images brought improvements in clinical data in terms of lesions definition and heterogeneity characterization. However at this stage only qualitative results are available for patient data.
Conclusions: The incorporation of functional and anatomical information in a synergistic fashion generates PET images of improved quality that might be a more robust starting point for tumour staging and characterization.
Methods: The proposed method takes advantage of the multiresolution property of the wavelet transform that is therefore applied to both functional and structural images. The correspondent resolution components of PET and CT are then combined after appropriate scaling and weighting. Although the method was validated for PET/MRI brain data we demonstrated the potentiality of our modified version with both phantom and patient experiment. We additionally compared the method with a standard resolution recovery approach based on point-spread-function reconstruction.
Results: The activity quantification and volume definition on phantom spheres showed better performance of our methodology compared to the clinical standard and the point-spread-function reconstruction. The used of segmented CT images brought improvements in clinical data in terms of lesions definition and heterogeneity characterization. However at this stage only qualitative results are available for patient data.
Conclusions: The incorporation of functional and anatomical information in a synergistic fashion generates PET images of improved quality that might be a more robust starting point for tumour staging and characterization.
| Original language | English |
|---|---|
| Title of host publication | 2014 Nuclear Science Symposium and Medical Imaging Conference Conference |
| Publication status | Published - 1 Dec 2014 |
| Event | IEEE Nuclear Science Symposium & Medical Imaging Conference 2014 - Seattle, United States Duration: 8 Nov 2014 → 15 Nov 2014 |
Conference
| Conference | IEEE Nuclear Science Symposium & Medical Imaging Conference 2014 |
|---|---|
| Country/Territory | United States |
| City | Seattle |
| Period | 8/11/2014 → 15/11/2014 |